Loose frequency information when storing with table=True to HDF5

[tobinvanpelt]

I have found what possibly is a bug. When storing data as HDF5 with table=True the read value of series.index.freq from HDF5 is set to None rather than its original values.

Please see: http://stackoverflow.com/questions/16311045/in-pandas-why-is-frequency-information-lost-when-storing-in-hdf5-as-a-table

Not sure if there is a limitation for some reason in PyTables or if this is a bug.

[jreback]

this is not implemented
as a workaround u can store it in an attribute
see the last cookbook entry: http://pandas.pydata.org/pandas-docs/dev/cookbook.html#hdfstore

I'll leave this issue open as a reminder - this is pretty straightforward if u would like to do a PR?

[jreback]

this is closed by #3531